Abstract
This paper considers phase-dependent probabilistic failure competitions in reliability analysis of wireless sensor network (WSN) applications. Particularly, a WSN can experience probabilistic functional dependence (PFD) behavior where operations of some sensors (referred to as probabilistic-dependent components) rely on functions of relay nodes (referred to as trigger components) with certain probabilities. Failure competitions exist in the time-domain between the local failure (e.g., transmission unit failure) of a relay node and propagated failures of its dependent sensors (e.g., jamming attacks), causing distinct system statuses. Further, a WSN may be subject to phased-mission requirements, incurring dynamics in system configuration and component failure behavior, as well as statistical dependencies across phases for a given component, which makes the reliability analysis more challenging. This paper models effects of phase-dependent probabilistic competing failures, and suggests a multi-valued decision diagram (MDD) based combinatorial procedure for reliability analysis of non-repairable WSN systems. The proposed method is applicable to arbitrary types of time-to-failure distributions for WSN components and probabilistic isolation factors. A case study of a WSN system for lab condition monitoring is presented to illustrate application and advantages of the proposed method.